Page 1 of 25

Achieving the Most Stringent CO2 Commercial Truck

Standards with Opposed Piston Engine

Dr. Gerhard Regner – Vice President, Performance and Emission

Fabien Redon - Vice President, Technology Development

John Koszewnik, Chief Technical Officer

Laurence Fromm – Vice President, Business and Strategy Development

Zoltan Bako – Director, Application Engineering

2

Abstract

With the passage of Euro 6, and the recent U.S. introduction of new CO2 limits for heavy-duty trucks

and buses, vehicle and engine manufacturers are facing a daunting challenge [1]. Compliance with these

regulations requires significant financial investments in new technologies, all designed to increase fuel

efficiency while decreasing emissions. But, to remain competitive, manufacturers cannot pass along

these costs to fleet owners.

One solution to this problem is the opposed-piston engine. This engine, which has been optimized by

Achates Power, was once widely used in a variety of applications including aviation, maritime and mili-

tary vehicles. After overcoming the architecture’s historical challenges, the Achates Power opposed-

piston engine now delivers a step-wise improvement in brake thermal efficiency over the most advanced

conventional four-stroke engines. In addition, with the elimination of parts such as the cylinder head

and valve train, it is also less complex and less costly to produce—making it even more appealing to

manufacturers.

After a brief overview of the opposed-piston architecture’s inherent efficiency benefits, this technical

paper features detailed performance and emissions results of a multi-cylinder Achates Power opposed-

piston engine configured to meet current commercial truck requirements. Presented for the first time in

Europe, these results demonstrate the engine’s ability to:

• Significantly improve fuel efficiency over the best diesel engines in the same class

• Comply with Euro 6/U.S. 2010 emissions standards

The discussion also includes an in-depth analysis of the opposed-piston, multi-cylinder test engine’s in-

dicated thermal efficiency, friction and pumping losses as well as a road map for achieving 47.6 percent

best-point brake thermal efficiency (BTE), which translates to 46.6 percent cycle-weighted BTE on me-

dium duty engine, while the same technology results 51.5% best point BTE, and 50.4% cycle weighted

BTE on a heavy duty engine. Furthermore, the technical paper provides a vibration analysis between

the Achates Power opposed-piston architecture and the inline six-cylinder, four-stroke engine, which

dominates the medium- and heavy-duty truck market.

Opposed-Piston Engine Architectural Advantages

Opposed-piston, two-stroke engines were conceived in the 1800s in Europe and subsequently developed

in multiple countries for a wide variety of applications, including aircraft, ships, tanks, trucks and loco-

motives. They maintained their presence throughout the twentieth century. An excellent summary of the

history of opposed-piston engines can be found in the SAE book, Opposed-Piston Engines: Evolution,

Use, and Future Applications by M. Flint and J.P. Pirault [2]. Produced initially for their manufactura-

bility and high power density, opposed-piston, two-stroke engines have demonstrated superior fuel effi-

ciency compared to their four-stroke counterparts. This section examines the underlying reasons for the

superior fuel efficiency and emissions. The OP2S diesel engine has the following efficiency advantages

compared to a conventional, four-stroke diesel engine:

Page 3 of 25

1. Reduced Heat Losses

The Achates Power opposed-piston engine, which includes two pistons facing each other in the same

cylinder, offers the opportunity to combine the stroke of both pistons to increase the effective stroke-to-

bore ratio of the cylinder working volume.

For example, when coupling two piston trains from a conventional, single-piston engine with a stroke-

to-bore ratio of 1.1, the resulting opposed-piston engine bore-to-stroke ratio is twice or 2.2. This can be

accomplished while preserving the engine and piston speed of the base design.

To achieve the same stroke-to-bore ratio with a single-piston engine, the mean piston speed would dou-

ble for the same engine speed. This would severely limit the engine speed range and, therefore, the

power output.

The increase in stroke-to-bore ratio has a direct mathematical relationship to the area-to-volume ratio of

the combustion space. For example, when comparing a single-piston engine to an opposed-piston engine

with the same piston and crank dimensions, the following outcome can be seen:

Table 1 OP2S compared to a single-piston engine.

In this example, the reduction in the surface area top volume ratio is a very significant 36%. The lower

surface area directly leads to a reduction in heat transfer.

Figure 1 Surface-to-volume ratio versus engine displacement for

an OP2S and conventional engine.

Single Piston OP2S

Trapped Volume/Cyl. 1.0L 1.6L

Bore 102.6 mm 102.6 mm

Total Stroke 112.9 mm 224.2 mm

Stroke-to-Bore Ratio 1.1 2.2

Compression Ratio 15:01 15:01

Surface Area (Min Vol.) 20 cm2

20 cm2

Volume (Min Vol.) 71 cm3

114 cm3

Area-to-Volume Ratio 0.28 0.18

4

Figure 1shows that the area-to-volume ratio of a six-liter, opposed-piston engine is equivalent to a 15-

liter, conventional diesel engine. This reduction in area-to-volume ratio is one of the main reasons why

larger displacement engines are more efficient than smaller ones. With the Achates Power opposed-

piston architecture, there is the opportunity to achieve the efficiency of much larger engines.

An additional benefit of the reduced heat losses in the opposed-piston engine, especially for commercial

vehicles, is the reduction in fan power and radiator size, further contributing to vehicle level fuel sav-

ings.

2. Leaner Combustion

When configuring an opposed-piston, two-stroke engine of the same displacement as a four-stroke en-

gine –f or example, converting a six-cylinder, conventional engine into a three-cylinder, opposed-piston

engine – the power that each cylinder has to deliver is the same. The opposed-piston engine fires each of

the three cylinders at each revolution while the four-stroke engine fires each of its six cylinders one out

of two revolutions.

Therefore, the amount of fuel injected for each combustion event is similar, but the cylinder volume is

more than 50% greater for the Achates Power opposed-piston engine. So for the same boost conditions,

the opposed-piston engine will achieve leaner combustion, which increases the ratio of specific heat. In-

creasing the ratio of specific heat increases the pressure rise during combustion and increases the work

extraction per unit of volume expansion during the expansion stroke.

3. Faster and Earlier Combustion at the Same Pres-

sure Rise Rate

The larger combustion volume for the given amount

of energy released also enables shorter combustion

duration while preserving the same maximum pres-

sure rise rate. The faster combustion improves ther-

mal efficiency by reaching a condition closer to con-

stant volume combustion. The lower heat losses as

described above lead to a 50% burn location closer

to the minimum volume. Figure 2 illustrates how the

heat release rate compares between a four-stroke en-

gine and the Achates Power opposed-piston engine.

The ideal combustion should occur at the minimum

volume and be instantaneous. The opposed-piston engine is much closer to this ideal condition at the

same pressure rise rate.

Ideal Engine Efficiency

1

11

c

idealr

rc = compression ratio

= ratio of specific heats

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0

100

200

300

400

500

600

700

800

900

1000

MFB

(-)

HR

R (J

/deg

)

Crank Angle (deg aMV)

4S Engine

OP2S Engine

Figure 2: Heat release rate comparison between a four

stroke and the OP2S.

Page 5 of 25

The aforementioned fundamental OP2S thermal efficiency advantages [3] are further amplified by:

– Lower heat loss due to higher wall temperature of the two piston crowns compared to a cylinder

head. (Reduced temperature delta).

– Reduced pumping work thanks to uniflow scavenging with the OP2S architecture giving higher ef-

fective flow area than a comparable four-stroke or a single-piston, two-stroke uniflow or loop-

scavenged engine [4].

– A decoupled pumping process from the piston motion due to the two-stroke architecture allows

alignment of the engine operation with a maximum compressor efficiency line [5].

– Lower NOx characteristics as a result of lower BMEP requirements because of the two-stroke cycle

operation [6].

Efficiency and Emissions Enablers

Combustion System

Achates Power has developed a proprietary combustion system [7] composed of two identical pistons

coming together to form an elongated ellipsoidal combustion volume where the injectors are located at

the end of the long axis [8] (Figure 3).

This combustion system allows:

– High turbulence, mixing and air utilization with both

swirl and tumble charge motion as is illustrated below with

the high turbulent kinetic energy available at the time of

auto ignition

– Ellipsoidal combustion chamber resulting in air en-

trainment into the spray plumes from two sides

– Inter-digitated, mid-cylinder penetration of fuel plumes

enabling larger λ=1 iso-surfaces

– Excellent control at lower fuel flow rates because of

two small injectors instead of a single higher flow rate

– Multiple injection events and optimization flexibility with strategies such as injector staggering and

rate-shaping [8]

The result is no direct fuel spray impingement on the piston walls and minimal flame-wall interaction

during combustion. This improves performance and emissions [9] with fewer hot spots on the piston

surfaces to further reduce heat losses [8].

Figure 3: Schematic of the combustion system with

plumes coming out of two side-mounted injectors

6

Air System

To provide a sufficient amount of air for combustion, two-stroke engines need to maintain an appropri-

ate pressure difference between the intake and exhaust ports (i.e. to scavenge exhaust out of the cylinder

after combustion and push in fresh air mass).

For applications that require the engine to change speed

and load in a transient manner, such as automotive applica-

tions, external means of air pumping are required. Among

the various possible configurations of the air system with

turbocharger and supercharger combinations, the layout as

described in Figure 4 is the preferred configuration [10].

Advantages of such an air system are summarized as fol-

lows:

– The compressor provides high pressure before the su-

percharger, which is multiplied by the supercharger.

This means low supercharger pressure ratios are suffi-

cient for high intake manifold density, reducing pump-

ing work.

– The maximum required compressor pressure ratio is

lower compared to regular turbo-only air systems of

four-stroke engines.

– The use of a supercharger recirculation valve allows greater control of the flow through the engine,

thus providing flexibility for precise control of boost, scavenging ratio, and trapped residuals to min-

imize pumping work and NOx formation across the engine map

– Lowering the flow through the engine by decreasing the pressure difference across the engine reduc-

es the pumping penalty at low load points. This, together with having no dedicated intake and ex-

haust stroke for moving mass from and to the cylinder improves BSFC.

– The supercharger and recirculation valve improves transient response [11].

– Accurate control of the engine pressure differential provides very good cold start and catalyst light

off capabilities [12] . For similar reasons, exhaust gas temperatures and catalyst light-off can be

maintained during low load and idle conditions.

– Low-speed torque is increased by selecting the appropriate gear ratios on the supercharger [9].

– Drive EGR with a supercharger reduces the required pumping work [9].

– Cool air and EGR together reduces fouling of the coolers [9][13].

Figure 4: Opposed-piston, two-stroke preferred air sys-

tem layout.

Page 7 of 25

Multi-Cylinder Research Engine Description

OP Engine Engineering Challenges

Historically, two-stroke, opposed-piston engines are known to have fuel efficiency advantages, but have

faced several engineering challenges that have kept them from going mainstream. The Achates Power

created a robust engine that satisfies the performance, emissions and durability standards of the 21st cen-

tury. The primary challenges that Achates Power had to overcome include finding an effective way to

reduce oil consumption, increase piston compression ring life, manage the thermal loads on the piston

and liner, and support 200+ bar cylinder pressures at the wrist pin.

The oil control strategy in a two-stroke engine is different than in a four-stroke due to the ports in the

cylinder liner, which also impact the piston ring wear. If there is not enough lubricant on the liner, the

ring life deteriorates. If there is too much oil, consumption increases. Two-stroke engine has a firing

event every crankshaft revolution whereas a four-stroke has a firing event every two revolutions. Inher-

ently, the two-stroke lacks the intake stroke which, for a four-stroke engine, allows for additional cool-

ing of the piston and cylinder liner. Creative solutions are required to sufficiently cool the piston and

cylinder liner. Traditionally, two-strokes have had limitations with wrist pin life at peak cylinder pres-

sures above 150 bar. This again is primarily driven by the lack of an intake stroke where inertia over-

comes the cylinder pressure and lifts the piston from the wrist pin and creating a void to be filled with

oil.

Single-Cylinder Development Engine

After focusing on the optimal engine architecture, Achates Power developed its single-cylinder variant,

designated the A48-1. These single-cylinder engines have been used for performance and emissions de-

velopment and have provided a platform for mechanical system technology development.

Achates Power utilized creative, but proven, solutions to overcome the presented engineering challeng-

es. In the case of oil consumption and ring life the focus was on liner honing techniques, piston ring ma-

terial and coating. This resulted in oil consumption that is on par with four-stroke engines in the medi-

um- and heavy-duty industry.

Improving liner and piston thermal management required a combined effort balancing heat in and out of

the liner and piston. On the hot side, combustion variables must be controlled while care must be taken

to avoid hot spots from the fuel plume flame fronts. The cold side of both the liner and piston uses tar-

geted cooling solutions to cool critical areas. Achates Power has utilized its proprietary real-time piston

and liner temperature measurement system to gain a fundamental understanding and control of thermal

issues.

Overcoming the 150 bar peak cylinder pressure limit of the typical two-stroke was accomplished by in-

troducing the bi-axial wrist pin. This offset bearing is fixed to the connecting rod, which forces the op-

posing journals to lift as it articulates. This has successfully allowed Achates Power to achieve 220 bar

peak cylinder pressure.

8

The combustion has been optimized for both fuel efficiency and emissions. Achates Power utilizes

unique combustion bowl shapes that allow for optimal mixing and scavenging by inducing additional

tumble in the combustion chamber. The piston shapes were designed as a system with the fuel injectors,

cylinder ports, crankshaft-to-crankshaft phasing and compression ratio.

After resolving these engineering challenges and achieving industry-leading fuel efficiency based on the

single-cylinder testing, it was time to prove these results carry over into a multi-cylinder design. Up un-

til this point, simulation and computational models were used to transfer results from a single-cylinder

to a multi-cylinder. Missing were cylinder-to-cylinder interactions with the air charge system and the

scaling of overall engine friction. At this point, Achates Power designed and built the three-cylinder

A48-3-16.

Multi-Cylinder Modular Development Engine

The A48-3-16 shares most of the power cylinder with the A48-1 and in an effort to reduce the develop-

ment schedule, many components are compatible. Similar to the A48-1, the A48-3-16 is designed for a

peak cylinder pressure of 200 bar with overload conditions of 220 bar. The block was cast from com-

pacted graphite iron (CGI).

The A48-1 was oriented with the cylinder axis in the horizontal plane while the A48-3-16 is oriented

vertically. The drive toward a vertical engine is based on customers’ preferences for packaging in a ve-

hicle.

The A48-1 and A48-3-16 engines were cre-

ated as a research test bed to quickly iterate

through multiple different designs. In creat-

ing such a platform, some compromises were

made versus how a production engine would

be conceived. Some examples of the experi-

mental aspects include:

● Higher overall engine mass – robust-

ness and quick turn around

● Larger package size – modu-

lar/swappable components

● Off-the-self air system components –

supercharger, turbocharger and cool-

ers, not tuned for the engine

● Higher friction

o Oversized off-the-shelf con-

necting rod big end and main

bearings

Figure 5: A48-3-16 Front and Rear View

Page 9 of 25

o Aftermarket oil and coolant pumps

o Remote mounted gearbox with redundant bearings causing over constraint

o Dual dry sump scavenging pumps and air-oil separators

● Modular gearbox connecting the exhaust and intake crank

● Modular FEAD

● Modular accessories

Friction can be trimmed in several areas. Due to available bearing sizes and the need for a robust devel-

opment platform, the loading calculations for both the connecting rod big end and main bearings result-

ed in oversized components. More of the cooling in the opposed-piston engine is done with the oil so

the efficiency of the oil pump is important and can be improved with deeper supplier involvement. The

gearbox, which connects the intake and exhaust crankshaft, is carried over and compatible with the sin-

gle cylinder A48-1. This gearbox was also designed with significant margins for robustness at the ex-

pense of friction and fuel efficiency.

Purpose-Built Multi-Cylinder

Many of these aspects of the engine allow for quicker and more productive development cycles. With a

purpose-built engine and known boundary conditions, the fuel efficiency can be further improved. By

removing the need for modular systems, the mass and packaging space can be dramatically improved.

The learnings from this modular A48-3-16 are used to design the next-generation application-specific

engines for volume production.

Engine Power and Torque Targets

The engine was configured to meet the following torque curve. The 4.9L three-cylinder engine has a

peak power output of 275hp@2200 RPM and a peak torque of 1100Nm from 1200 to 1600 RPM (Table

3)

The air and EGR system was sized to achieve EGR and air-fuel ratio levels suitable to comply with U.S.

2010 emissions levels when coupled with conventional SCR and DPF aftertreatment.

Figure 6: A48-3-16 engine specification and power and torque curves

Displacement 4.9 L

Arrangement, # of Cyl. Inline 3

Bore 98.4 mm

Total Stroke 215.9 mm

Stroke-to-Bore Ratio 2.2

Compression Ratio 15:01

Nominal Power (kW@rpm) 205 @ 2200

Max. Torque (Nm@rpm) 1100 @ 1200-1600

A48-3-16 engine specification

10

Engine Build

The engine build team worked for approximately two months to deliver the A48-3-16engine to the test

group. During this time, all components were verified for form, fit and function. Critical dimensions and

clearances were recorded for use during engine inspections and to monitor component wear. To meet

the aggressive schedule, the procurement of components was planned so assembly could begin while

waiting for subsequent systems. The power cylinder and block were assembled first, followed by the

fuel, coolant, air charge systems and finally the FEAD. The engine was then instrumented and connect-

ed to the test cell. The engine was first run on fuel eleven months after the initiation of the project.

Engine Testing - Instrumentation

In-cylinder pressure is measured at 0.5° crank-angle intervals with three AVL GH14D Select piezoelec-

tric pressure transducers coupled to Kistler 5064 charge amplifiers. The cylinder pressure signal is

pegged to an average of the intake and exhaust manifold pressures during scavenging, measured with

Kistler 4005B and 4045A high-speed pressure transducers, respectively. Custom in-house software is

used to acquire and process the crank-angle based data.

Exhaust emissions are measured with an FTIR in conjunction with a California Analytical Instruments

(CAI) emissions analyzer. These are used to measure the steady-state concentration of five exhaust spe-

cies (CO2, CO, O2, HC, NOx) and intake CO2. An AVL 483 Micro Soot Sensor provides a measure of

exhaust soot content in real time. A Davinci DALOC is used for real-time oil consumption measure-

ment.

Torque is measured with a Kistler 4504B Torque Flange with a capacity of 2000 Nm and an accuracy of

± 0.05%. The torque flange is mounted in the driveline between the engine and the dyno absorber.

The Re-Sol RS 515A-125 Fuel Flow Measurement System utilizes a “float tank”-style level controller

to combine the return fuel with the incoming fuel and reduce measurement to a single flow path. The

measurement is done with a Micro Motion CMFS010 with an accuracy of ± 0.05% and a capacity of

Figure 7: Engine components

Page 11 of 25

110 kg/hr. The test cell instrumentation is calibrated on a quarterly basis, with the exception of emis-

sions measurement, which is calibrated daily.

Performance and Emissions Test Results

The measured fuel consumption confirmed expectations for the development engine; combustion per-

formance and pumping losses were in good agreement with predictions. The best point fuel consump-

tion of 194.5g/kWh occurred at both A100 and B100 point, which equates 43.1% brake thermal effi-

ciency. The best indicated efficiency of 52.3% occurred at C25 operating point. The SET 12 mode

weighted average fuel condition equates to 201.1 g/kWh. The 12 mode cycle measured BSNOx averag-

es 3.30 g/kWh enabling tailpipe emission compliance US 2010 with typical SCR conversion efficiency.

The NOx map (Figure 9) shows noticeable dependency with engine speed, which is a consequence of

the speed dependency of the residual gas content during the uniflow scavenging process.

The 0.050g/kWh weighted average 12mode results of BS Soot (Figure 10) is in a good range not only to

meet tailpipe emissions but also to achieve low particulate filter regeneration frequency. Slightly higher

BSSoot emission is measured around C100 due to lower air mass than desired, future engine build are

Figure 8: Multi-cylinder BSFC map Figure 9: Multi-cylinder BSNOx map

Figure 10: Multi-cylinder BSSoot (AVL415S) map Figure 11: Multi-cylinder BSHC map

12

expected to perform better at that area.

The BSHC (shows in Figure 11) weighted average of 0.06 g/kWh is extremely low, does not even re-

quire any aftertreatment reduction.

Table below summaries the expected tailpipe emission with corresponding aftertreatment efficiencies.

EPA 2010

Tailpipe limit

[g/kWh]

Engine out meas-

urement [g/kWh]

Aftertreatment

conversion effi-

ciency

Expected tailpipe

emissions [g/kWh]

NOx 0.27 3.30 93% (SCR) 0.23

PM 0.013 0.050 99% (DPF) 0.001

HC 0.19 0.057 99% (DOC) 0.001

The low measured BSCO (shown in Figure 12) values across the entire map confirm the excellent quali-

ty of the combustion process of the proprietary Achates Power combustion system.

Figure 12: Multi-cylinder BSCO map Figure 13: Multi-cylinder ITE map

Figure 14: Multi-cylinder pumping loss map Figure 15: Multi-cylinder friction map

Page 13 of 25

Figure 16: Turbocharger compressor (left) and supercharger map with operating points

Figure 16 shows the operating point on the turbocharger and supercharger compressor maps. Unlike

typical four-stroke engines, the OP2S engine does not need as wide of a compressor map. Therefore,

achieving good efficiency at rated power while still maintaining a good surge margin at peak torque and

other low-speed, high-load points was not as difficult

Table 2 Selected SET 12 mode measurement data

Engine Condition A25 A75 A100 B50 B75 B100 C25 C75 C100

Engine Speed rpm 1400 1400 1400 1800 1800 1800 2200 2200 2200

IMEP bar 4.6 12.3 15.9 8.1 11.5 15.1 4.3 10.8 14.3

BMEP bar 3.5 10.5 13.9 6.5 9.7 13 2.8 8.6 11.6

Indicated Power kW 52.7 141.2 182.3 120 169.5 222.5 77.1 194.2 258.2

Brake Power kW 40.2 120.3 160.2 95.8 143.5 191.8 51.4 154.8 209.2

Indicated Thermal Efficiency %fuel 51.6 50.3 49.2 52.1 50.9 50.1 52.3 52.01 51.46

Brake Thermal Efficiency %fuel 39.3 42.8 43.1 41.5 43 43.1 34.8 41.4 41.6

Friction Loss %fuel 10.4 5.1 4.0 8.3 6.3 5.3 13.7 7.8 7.1

Pumping Loss %fuel 1.9 2.4 2.1 2.3 1.6 1.7 3.8 2.8 2.7

ISFC  (Engine) g/kWh 162.8 166.9 170.8 161.2 164.8 167.6 160.6 161.5 163.2

BSFC  (Engine) g/kWh 213.4 195.8 194.5 201.9 194.7 194.5 240.7 202.6 201.4

BSNOx g/kWh 2.336 4.743 4.789 3.853 4.299 3.351 2.288 1.841 1.037

BSSOOT g/kWh 0.013 0.011 0.036 0.017 0.017 0.05 0.032 2.276 1.218

BSCO g/kWh 0.398 0.286 1.067 0.224 0.384 1.173 0.527 0.325 1.154

BSHC g/kWh 0.071 0.038 0.038 0.056 0.054 0.063 0.094 0.062 0.058

Peak Cylinder Pressure bar 79 162 198 129 161 197 90 173 200

50% Mass Burned Fraction deg aMV 2.7 3.1 3 2 2.7 4 0.3 1.6 4.6

Burn Duration 10-90% deg 15.7 27.8 31.1 21.4 22.6 25.0 18.7 21.8 22.0

Air/Fuel Ratio - 30.2 27.3 24.4 28.9 23.9 22.1 30.9 24.4 21.3

External EGR Rate % 32.3 31.5 30.2 27.6 25.6 26.0 32.0 33.2 34.0

Intake Manifold Pressure bar 1.338 2.297 2.705 2.034 2.420 2.989 1.593 2.730 3.305

Intake Manifold Temperature degC 38 40 43 40 40 44 39 40 45

Turbine Outlet Temperature degC 250 278 311 264 323 343 240 287 327

14

Multi-Cylinder Engine Modelling

The first portion of this paper has discussed initial results from a multi-cylinder engine built for research

and development purposes. A production intent 4.9L engine targeted for 2017+ release will include im-

provements in combustion, pumping and friction.

GT-Power, a capable one-dimensional engine and vehicle simulation tool, is used to predict what the

performance of such an engine would be.

A two-step approach is used in making future multi-cylinder predictions from current 4.9L multi-

cylinder measurements. The first step is a correlating the 1-D model to test cell measurements. This al-

lows for accurate determination of in-cylinder heat transfer, trapped composition, friction losses and

pumping losses. Step two involves assessing what model changes and assumptions will be necessary to

make the proper prediction on 4.9L engine, than apply those on a 11L engine to predict a heavy duty

engine performance.

The 1-D model requires a detailed characterization of the scavenging process because it is important to

arrive at the correct concentrations of fresh air and residual gas in the cylinder prior to the start of the

closed-cycle portion of the simulation. For this reason, the scavenging efficiency was measured in the

test engine using an in-cylinder CO2 sampling method, and the scavenging efficiency versus scavenge

ratio relationship was used in both the correlation and prediction process.

Before accurate correlation can begin, the friction and engine accessory efficiencies are measured with

dedicated tests and then input into the 1-D model. The multi-cylinder model air-handling system con-

sists of a supercharger, a turbocharger, a charge air cooler after each compression stage and EGR cool-

ers. The size and characteristics of the air-handling system components are application specific. The

compressor and turbine is modeled using map data provided by a turbocharger supplier, and the super-

charger model uses a full map obtained from a supercharger supplier.

In the correlation model, the combustion chamber geometry, the piston motion and the porting profiles

are identical to what exists in the multi-cylinder engine. Engine speed, fuel flow rate, air flow rate, EGR

percentage, cylinder pressures at 30° before minimum volume, brake torque, and the intake/exhaust

manifold and compressor inlet/turbine outlet pressures and temperatures match the measured values.

The rate of heat release is derived from the measured cylinder pressure and is input directly into the

combustion sub-model. From this, the trapped conditions in the cylinder are determined as well as the

in-cylinder heat transfer coefficient using the measured indicated thermal efficiency as a target. The

model is iterated until cylinder pressure traces, crank angle resolved intake and exhaust pressures, air

system pressure drops and temperatures, and turbo machinery performance matches the measured val-

ues. Once the model has been fully correlated over the entire operating range, it is then transformed into

the prediction model. This is accomplished by fixing the in-cylinder heat transfer coefficient and con-

trolling fuel flow with a brake mean effective pressure (BMEP) controller. Hardware changes, identified

improvements in component performance, or certain calibration changes can be applied to this model to

predict how overall engine performance and losses change as a result of the given change. Modifications

can be performed one at a time to determine individual performance gains or all at once to find the ef-

fects of a full engine upgrade.

Page 15 of 25

Model Correlation

3D Gas Exchange Modeling

CFD model of the A48-3-16 engine air system was created to model the gas exchange process. The do-

main starts after the supercharge bypass and ends at the EGR pickup and turbo inlet. The model con-

tains six moving pistons that cover and uncover the intake and exhaust ports and intake and exhaust

manifolds. Upstream and downstream pipes were also included to accurately predict pressure wave dy-

namics in the system. The simulations were run using CONVERGE 2.1 with moving boundaries.

Boundary conditions are obtained from 1D GT-Power complete engine model simulations at different

engine loads and speeds. The simulation tracks residual gases and fresh air charge in all three cylinders

to predict Delivery Ratio (DR), Trapping Efficiency (TE), Scavenging Efficiency (SE) and Charging Ef-

ficiency (CE).

The A48-3-16 engine is equipped with high-speed pressure measurements at the intake and exhaust

manifolds. A hot wire anemometer and a laminar flow element are used to measure mass flow rate per

cycle in the engine at the test cell. With these measurements it is possible to validate the CFD simula-

tion results.

Figure 18 begins at 120° crank angle, which is minimum volume for Cylinder 2. The comparison shows

very good correlation between predicted and measured pressure waves in both intake and exhaust mani-

folds.

Correct prediction of the pressure wave dynamics is essential to have a good delivered mass correlation.

At 1800 rpm and a 50% load point, the measured delivered mass per cycle was 9043.3 mg/cycle where

CFD simulation predicts 8876.7 mg/cycle. The difference in measured and predicted delivered mass per

cycle is below 2%.

Good correlation is critical to build confidence in the simulation methodology and results. With this

confidence, many virtual prototypes can be tested in a short period of time to help choose an optimal de-

sign early in the engine design phases.

Figure 17: Intake and exhaust pressure correlation

16

3D Combustion System Modeling

Commercially available CONVERGE CFD software [14, 15] is used to perform in-cylinder simulations

of the OP2S combustion system. CONVERGE offers detailed chemistry solvers along with hydro-

dynamic solvers that provide a basis of combustion simulation. The methodology, adopted for in-

cylinder combustion simulation, is a combination of multi-cylinder engine (MCE) air flow simulation as

described in the previous sub-section and coupling the MCE air flow results with in-cylinder combus-

tion simulation for only one cylinder. In the present study, the cylinder towards the end of the intake

manifold (namely Cylinder 3) is chosen as a representative cylinder for simulation purposes. Because of

its extreme location away from the inlet of the intake manifold, Cylinder 3 is a good choice of represen-

tation to improve combustion using hardware changes, such as injector hole size, spray patterns, port

orientation, etc. For any hardware changes in Cylinder 3 that result in improved combustion, similar

improvements are also observed in other cylinders. Intake and exhaust port geometries are not included

as only the closed portion of the cycle from exhaust port closing (EPC) (= -118 degree aMV) to exhaust

port opening (=114.5 degree aMV) is simulated for the purpose of combustion analysis. This is compu-

tationally efficient compared to simulating an entire multi-cylinder with detailed chemistry solvers. The

initial conditions for the simulation require trapped thermodynamic conditions and trapped flow condi-

tions. Trapped flow conditions include the velocity field, turbulent kinetic energy and dissipation rate,

which are obtained from the entire MCE gas-exchange simulation as described in the previous sub-

section. The trapped pressure is specified based on the cylinder pressure measurements, and the trapped

composition and temperature are obtained from a two-zone mixing model based on scavenging meas-

urements. Note that trapped composition and temperature can also be obtained using well-correlated,

one-dimensional code, such as GT-Power, that simulates a multi-cylinder engine with imposed combus-

tion characteristics. In the case of the Achates Power engine, the scavenging measurements provide di-

rect means of deriving trapped composition and temperature.

A detailed chemistry model involving a well-known reduced chemistry mechanism for n-heptane (diesel

fuel surrogate) with 35 species and 77 reaction steps [16], which include a NOx sub-mechanism, is

used. Soot emissions are modeled using a two-step model, which includes a Hiroyasu formation step

with acetylene as the precursor [17], and an oxidation step involving carbon oxidation by O2 molecules

Figure 18: Comparison of measured and model cylin-

der pressure curves at B75

Page 17 of 25

[18]. Sprays are modeled using a modified KH-RT break-up model without the use of an ad-hoc

breakup length [15, 19] and the O’Rourke collision model [15, 20], whereas turbulence is modeled us-

ing the RNG k-ε model [15, 21]. Fuel injection rate profiles are specified based on measured data from a

state-of-the-art, in-house fuel laboratory with IFR (Injection Flow and Rate) capabilities [22]. Example

of the archived correlation at B75 point is shown in Figure 19.

Performance and Emissions Roadmap

The Achates Power opposed-piston, two-stroke engine has demonstrated a cycle weighted average

brake thermal efficiency (BTE) of 41.8% with engine hardware and calibration that are still in an early

stage of development. Higher engine thermal efficiencies will be achieved through hardware and cali-

bration improvements, some of which are unique to the Achates Power engine architecture and some of

which are industry-wide advancements. To quantify the effect of these possible improvements, a BTE

roadmap has been developed. The potential efficiency improvements are estimated based on internal

analysis as well as findings from a report on fuel economy technologies to the United States National

Academy of Sciences. Figure 19 shows the energy balance and efficiency improvements using the 12-

mode weighted average results. The predicted brake-specific fuel consumption map is also shown in

Figure 20.

In order to achieve a cycle-weighted brake thermal efficiency of 46.6% or 180g/kWh BSFC, the indi-

cated closed-cycle efficiency, pumping work, mechanical friction and the power consumption of the en-

gine accessories all require further improvements. A detailed discussion of each improvement oppor-

tunity can be found in the following sections to support the quantitative estimates put forth in Figure 19.

Combustion improvements

The ability to convert fuel energy to mechanical energy efficiently and cleanly while still meeting exter-

nal mechanical and emission constraints is paramount to a successful internal combustion engine. For

the current weighted average, a gross indicated thermal efficiency of 51.0% is achieved using a calibra-

tion with a maximum of 10 bar/deg maximum pressure rise rate, and a greater than 90% efficient SCR

device that allows 2010 U.S. emissions requirements to be met with 3.3 g/kW-hr engine-out NOx.

Engine Speed (RPM)

To

rqu

e (

N-m

)

Preliminary A48-3-16 2017 Roadmap BSFC

1200 1400 1600 1800 2000 2200220

300

400

500

600

700

800

900

1000

1096

BS

FC

(g/k

W-h

r)

160

170

180

190

200

210

220

230

240

250

260

Figure 201: A48-3-16 brake-specific fuel consumption roadmap

prediction for 2017+

Figure 19: Energy balance and efficiency improvements of cycle-

weighted average

Fuel Energy 100%

Indicated Closed - Cycle Efficiency

51.0% 51.8%

Coolant & Exhaust Heat Rejection

49.0% 48.2%

Pumping

2.0% 0.7%

Friction and

7.2% 4.5%

Brake Power

41.8% 46.6% Engine Accessories

18

Improving the combustion system is a primary step toward increasing the brake thermal efficiency. In

the present analysis, the effects of changes to the designed engine hardware and calibration on indicated

thermal efficiency were evaluated using CFD results, new piston bowl shapes were developed, fuel in-

jector nozzle configurations were designed to more favorable heat transfer and combustion characteris-

tic. The cylinder port design also can be improve scavenging performance. The new piston bowl and

nozzle design resulted in a 0.7%fuel improvement over the current status.

Gas Exchange

Gas exchange losses for a two-stroke engine are represented by the pumping work provided by the

crankshaft-driven supercharger. The mechanical losses to drive the supercharger are included in the ac-

cessory power consumption and will be discussed in the next section. In general, the supercharger pow-

er requirements depend on the pressure losses of the entire air system, the additional pumping to com-

pensate for short-circuiting of fresh air during scavenging, the EGR rate and the efficiencies of the

supercharger and turbocharger. All of these effects were quantified by a 1D engine performance simula-

tion tool. The baseline is the measurement indicated pumping losses, equivalent to 2.0%fuel.

Several measures were applied in succession to reduce the pumping work. First, the porting arrange-

ment was optimized for the future hardware. Secondly the intake system was modified allowing more

efficient air induction, reducing pressure drop. These design differences were modeled in CFD to quan-

tify the scavenging improvements over the design currently under testing. Thirdly more efficient charge

air cooler were modeled, allowing the total system pressure to be reduced while increasing the trapped

air-fuel ratio. The 1D simulation model determined 0.6%fuel as a results of all these improvements.

As Achates Power moves to a 2017 production-intent engine configuration, supercharger suppliers will

be able to produce a unit that is better suited for the opposed-piston, two-stroke diesel engine as op-

posed to the high pressure ratio superchargers currently used on four-stroke engines. With current tech-

nology, the Achates Power opposed-piston engine utilizes a very small portion of the available super-

charger maps that is outside of the maximum efficiency island (see Figure 16, supercharger map). By

designing a unit that focuses on higher efficiency in the low speed, low pressure ratio portion of the

map, pumping work will be reduced. Finally, all of these advancements in pumping efficiency will pro-

Figure 21: Contributors of BTE improvements of cycle-weighted average

Page 19 of 25

vide the opportunity to match a different turbocharger. Higher exhaust temperatures, improved super-

charger pumping, and future improvements in turbocharger technology help to improve turbocharger ef-

ficiency. The improvement with a supercharger better suited for opposed-piston, two-stroke diesel and

better turbocharger matches provide0.5%fuel improvements.

All of these advances drop pumping losses down to 0.7%fuel as shown in Figure 19.

Friction and Engine Accessories

The power cylinder friction (ring/liner and piston/liner friction) closely matches the situation in a four-

stroke diesel engine, since both engine types employ a slider-crank mechanism. It is, therefore, a rea-

sonable strategy to leverage the same industry-wide advancements in the area of tribology and advanced

lubricants to lower the friction losses of the power cylinder in an opposed-piston engine. For this

roadmap, the friction reduction for the power cylinder, bearings and geartrain are projected to divide up

as follows: 0.35%fuel are gained from the power cylinder based on further optimization of the ring and

piston skirt contours combined with advanced surface textures and/or coatings; 0.87%fuel from the main

and rod bearings based on size optimization and oil temperature management; and 0.94%fuel from the

geartrain based on optimized geartrain design and lastly the lower lube pump power requirement allows

0.84%fuel reduction. Combined, the projected improvements is 3%fuel, which highlights the improvement

potential starting from a robust oversized design such as Achates Power A48-3-16 development engine.

It is important to note that some previously introduced improvement actually increase the friction, hence

more reduction is required to achieve 46.6% cycle weighted BTE than Figure 19 suggests.

By incorporating all friction reduction measures, the friction losses reduce to 4.5%fuel.

Heavy Duty Engine Prediction

Previous section detailed the currently achieved engine performance and emission results, as well the

road map to achieve 46.6% weighted cycle average brake thermal efficiency representative of a volume

production medium duty engine design. Using the 4.9L multi cylinder engine correlated models a heavy

duty engine model was created.

Table 3: Heavy duty engine specification

Figure 22: Heavy duty engine power and torque curve

Displacement 11.0 L

Arrangement, # of Cyl. Inline 3

Bore 125 mm

Total Stroke 300 mm

Stroke-to-Bore Ratio 2.4

Compression Ratio 15:01

Nominal Power (kW@rpm) 390 @ 1700-2100

Max. Torque (Nm@rpm) 2200 @ 1200-1600

HD engine specification

20

Completive power and torque target were selected in this category (Table3 and Figure 22).

All improvement potential described in previous section was applied to create a BSFC map representa-

tive for much larger bore engine. The stroke to bore ratio was increased to 2.4 from 2.2 of the 4.9L en-

gine in order to improve area to volume ratio and also improve gas-exchange characteristics over the

4.9L engine. Engine calibration was assumed to reach similar engine out emission level as of 4.9L en-

gine. The combustion system was scaled from 98.4mm bore to 125mm bore size. The more favorable

area to volume ratio and the lower in-cylinder heat transfer allow reaching higher brake thermal effi-

ciencies shown in Figure 23; the prediction shows 51.5% best point brake thermal efficiency can be

achieved, which equates 162.7 g/kWh BSFC at A75 point. The SET 12 mode weighted cycle average

fuel consumption is calculated 166g/kWh, brake thermal efficiency of 50.5%.

Engine Vibration

The inherent vibration characteristics are an important consideration when evaluating engine architec-

tures for any on-road application. The current heavy- and medium-duty market is dominated by inline

six-cylinder, four-stroke engines. This baseline configuration features theoretically “perfect” force and

moment balancing, with the only residual effect being the reaction from the engine output torque.

Therefore, any un-cancelled residual forces or moments from the opposed-piston, two-stroke, three-

cylinder engine will be an additional input to the engine mount system design.

The engine output torque reaction moments will be comparable to an inline six-cylinder, four-stroke en-

gine with the same crank rotational speed and mean brake torque. This is because the frequency of the

firing events is the same between both cases. The other assumptions in this statement are that the rota-

tional inertias and peak cylinder pressures of the systems are comparable.

The opposed-piston architecture inherently balances out the majority of the piston acceleration forces

within each cylinder. As the intake side piston decelerates towards the injector plane, the exhaust side

piston also decelerates in a similar magnitude, but in the opposite direction. The only offset is a result of

the phase shift between the two pistons. The exhaust piston is phased slightly ahead of the intake piston

Achates Power Heavy-Duty OP2S

Figure 23: 11L Heavy duty engine BSFC map predicted for 2017+

Page 21 of 25

to maintain favorable intake to exhaust port time-areas and overall expansion ratio. As a result, there is a

small residual piston inertial force from each cylinder.

The firing order for a three-cylinder, opposed-piston, two-stroke (OP2S) features even 120° firing

events. When the residual force from one pair of pistons is at a maximum, the residual forces from the

other two pairs of pistons are half of the magnitude each, and in the opposite direction relative to the

first. This means that the forces effectively cancel out. To confirm this, a kinematic model was created

in CREO/Mechanism. The system analysed was the A48-3-16 research engine at an 8° exhaust crank

lead, operating at peak power.

The mechanism analysis was configured to output the residual forces and moments, neglecting compo-

nent compliance and system resonances. The residual block forces are shown in Figure 24. The magni-

tudes of these residual forces are exceptionally small, and may be neglected for any engine mount de-

sign.

Since the internal forces essentially cancel, and the torque reaction moments are comparable, the mo-

ments about the X and Y axes are the last excitations to consider. The output from the mechanism mod-

el resulted in the moments shown in Figure 25.

Moment magnitudes in this range are well within standard engine mount design capability. This charac-

teristic curve features both first- and third-order content. Since the first-order content magnitude exceeds

the third-order content, there is an opportunity to reduce the peak even below this reasonably low level.

Adding equal and opposite masses to the end of one crank (or any shaft rotating at crank speed) will

Figure 25: Residual unbalanced block moments at peak power

Figure 24: Residual block forces at peak power

22

counteract the moment about the X-axis without generating any residual forces. There will be a first or-

der sine wave added to the residual moments about the Y-axis as a result of this balancing.

Again, the CREO/Mechanism model was used to demonstrate this concept. The results are shown in

Figure 26.The mass and eccentricity of the “balancing” feature was increased until the peak magnitude

of the moments about the Y axis approached the peak magnitude of the moments about the X axis. Of

course, if the particular mounting system design isolates the moments in either the X- or Y-axis more

effectively than the other, this peak moment about the X-axis reduction technique may be adjusted.

The behaviour of the system trends increase the peak moments as a function of the square of the engine

speed, and as a linear function of the exhaust crank lead. For example, this 8° exhaust crank lead results

in a peak magnitude for the moment about X of less than 460 Nm. If the exhaust crank lead was reduced

to 6°, the peak magnitude for the moment about X would be less than 345 Nm at the same engine speed.

Figure 26: Residual balanced block moments at peak power

Page 23 of 25

Summary/Conclusions

● Achates Power is pleased to publish the first fully autonomous opposed-piston engine brake re-

sults, including fuel consumption and emissions.

● The performance demonstrated there is achieved with all the engine accessories and auxiliaries

driven by the engine and without applying the latest developments that would be applicable to

the opposed-piston engine, such as waste heat recovery, low friction coatings, thermal barrier

coatings, electrified accessories, two-stage turbochargers and turbo-compound.

● The technologies that Achates Power has developed for the opposed-piston engine have demon-

strated the ability to exceed any four-stroke engine of equivalent size.

● The measured results shown in this paper are from a very initial attempt at demonstrating multi-

cylinder brake performance. The significant learnings from this exercise will be the basis for

continued further improvements leading to a potential cycle average fuel economy over the 12-

mode points for 46.6% BTE on 4.9L engine size, while same technologies applied on 11.0L

heavy duty engine the potential cycle average fuel economy over the 12-mode points is 50.5%

BTE.

● This paper also describes how the Achates Power engine can be configured to be compact, light

and easy to integrate in a vehicle.

24

●

References

1. DieselNet, Diesel Exhaust Emissions Standards, Retrieved from http://www.dieselnet.com/standards/. 2. Flint, M. and Pirault, J.P., “Opposed Piston Engines: Evolution, Use, and Future Applications”, SAE International, Warrendale,

PA ISBN 978-0-7680-1800-4, 2009. 3. Herold, R., Wahl, M., Regner, G., Lemke, J., and Foster, D., “Thermodynamic Benefits of Opposed-Piston Two-Stroke En-

gines,” SAE Technical Paper 2011-01-2216, 2011, doi: 10.4271/2011-01-2216. 4. Regner, G., Naik, S., “Not All Two-Stroke Engines Are Created Equal”, Retrieved from http://www.achatespower.com/diesel-

engine-blog/2013/09/27/not-all-two-stroke-engines-are-created-equal/, 2013. 5. Regner, G., “Turbocharger Efficiency: An Underappreciated OP2S Advantage”, Retrieved from

http://www.achatespower.com/diesel-engine-blog/2013/01/23/turbocharger-efficiency/, 2013. 6. Regner, G., “The Achates Power Engine: Low NOx and Superior Efficiency”, Retrieved from

http://www.achatespower.com/diesel-engine-blog/2013/02/27/low-nox/, 2013. 7. Fuqua, K., Redon, F., Shen, H., Wahl, M., and Lenski, B., “Combustion Chamber Constructions for Opposed-Piston Engines”,

U.S. Patent Application US20110271932. 8. Venugopal, R., Abani, N., MacKenzie, R., “Effects of Injection Pattern Design on Piston Thermal Management in an Opposed-

Piston Two-Stroke Engine”, SAE International Technical Paper 2013-01-2423, 2013, doi:10.4271/2013-01-2423. 9. Regner, G., Fromm, L., Johnson, D., Koszewnik, J., Dion, E., Redon, F., “Modernizing the Opposed-Piston, Two-Stroke Engine

for Clean, Efficient Transportation”, SAE International Technical Paper 2013-26-0114, 2013, doi:10.4271/2013-26-0114. 10. Pohorelsky, L., Brynych, P., Macek, J., Vallaude, P., Ricaud, J., Obsernesser, P., Tribotté, P., “Air System Conception for a

Downsized Two-Stroke Diesel Engine”, SAE International Technical Paper 2012-01-0831, 2012, doi:10.4271/2012-01-0831. 11. Ostrowski, G., Neely, G., Chadwell, C., Mehta, D., Wetzel, P., “Downspeeding and Supercharging a Diesel Passenger Car for

Increased Fuel Economy”, SAE International Technical Paper 2012-01-0704, 2012. 12. Kalebjian, C., Redon, F., and Wahl, M. “Low Emissions and Rapid Catalyst Light-Off Capability for Upcoming Emissions Regu-

lations with an Opposed-Piston, Two-Stroke Diesel Engine”, Emissions 2012 Conference. 13. Teng, H. and Regner, G., “Characteristics of Soot Deposits in EGR Coolers”, SAE International Journal of Fuels and Lubri-

cants, Vol. 2, No. 2, pp. 81-90, 2010. Also published as SAE Technical Paper 2009-01-2671, 2009, doi:10.4271/2012-01-0704.

14. Senecal, P.K., Richards, K.J., Pomraning, E., Yang, T., Dai, M.Z., McDavid, R.M., Patterson, M.A., Hou, S., and Shethaji, T., “A New Parallel Cut-Cell Cartesian CFD Code for Rapid Grid Generation Applied to In-Cylinder Diesel Engine Simulations,” SAE Paper No. 2007-01-0159, 2007.

15. Richards, K.J., Senecal, P.K., Pomraning, E., “CONVERGE (Version 1.4.1),” Convergent Science Inc., Middleton, WI, 2012. 16. Patel, A., Kong, S., -C., and Reitz, R.D., “Development and Validation of a Reduced Reaction Mechanism for HCCI Engine

Simulations,” SAE Technical Paper 2004-01-0558, 2004. 17. Hiroyasu, H., and Kadota, T., “Models for Combustion and Formation of Nitric Oxide and Soot in DI Diesel Engines,” SAE Pa-

per No. 760129, 1976. 18. Nagle, J., and Strickland-Constable, R.F., “Oxidation of Carbon Between 1000-2000 C,” Proceedings of the Fifth Carbon Con-

ference, Vol. 1, p.154, 1962. 19. Patterson, M.A., “Modeling the Effects of Fuel Injection Characteristics on Diesel Combustion and Emissions,” Ph.D. Thesis,

University of Wisconsin-Madison, 1997. 20. O’Rourke, P.J., “Collective Drop Effects on Vaporizing Liquid Sprays,” Ph.D. Thesis, Princeton University, 1981. 21. Han, Z., and Reitz, R.D., “Turbulence Modeling of Internal Combustion Engines Using RNG k-ε Models,” Combustion Science

and Technology, Vol. 106, 1995. 22. Klyza, C., “Optical Measurement Methods used in Calibration and Validation of Modeled Injection Spray Characteristics,” Post-

er P7, presented in the 2010 Directions in Engine-Efficiency and Emissions Research (DEER) Conference. 23. Delgado, o and Lutsey, N., The U.S. SuperTruck program expediting the development of advanced heavy-duty vehicle effi-

ciency technologies, International Council on Clean Transportation, 2014

Contact Information

Gerhard Regner

Vice President, Performance and Emission

Achates Power, Inc.

Email Address: regner@achatespower.com

25